Every reliability conversation eventually arrives at the same quiet fact: the person who actually knows why Line 3 behaves the way it does is 58 years old, and has no successor.
Not "no trained successor." No successor at all. The req has been open for fourteen months. The last two candidates wanted remote work, which is difficult when the job is to put hands on a gearbox.
We have spent a decade talking about predictive maintenance, connected sensors, and AI-assisted diagnostics — and almost none of it about who will be on the floor to act on any of it. The technology roadmap assumes a workforce that is, right now, walking out the door faster than it is being replaced.
The scale of what's coming
The numbers are not subtle, and they are not a forecast for some distant decade. They describe the workforce on the floor today.
Deloitte and The Manufacturing Institute project that US manufacturing will need as many as 3.8 million new workers between 2024 and 2033 — and that roughly 1.9 million of those roles could go unfilled if the talent gap isn't closed. About half. And the hardest category to fill is not general labour; it is skilled technical work — the maintenance technicians, millwrights, and controls specialists who keep equipment running.
The demographic math underneath is the real story. More than one in five skilled-trades workers is over 55. In facilities and maintenance specifically, close to 39 per cent are over 55 and approaching retirement. The people who hold the institutional knowledge are, as a group, closer to leaving than almost any other part of the workforce. In Canada, the same wave arrives on the same schedule: about one-quarter of the manufacturing workforce is already 55 or older, with fewer than 10 per cent under the age of 25 to replace them.
Why this is a reliability problem, not just an HR problem
It is tempting to file this under recruiting and move on. That is a mistake, because the shortage lands directly on the metrics reliability teams are measured by.
A missing maintenance technician does not show up as an empty chair. It shows up as longer unplanned downtime, because the person who could diagnose the fault in twenty minutes is no longer there and the replacement takes two hours. It shows up as more reactive work and less preventive work, because a short-staffed team defends against breakdowns instead of preventing them. It shows up as slower commissioning of the very equipment that was supposed to modernise the plant. And it shows up as rising safety and compliance risk, because experience is what keeps people safe around energy sources that do not forgive mistakes.
Every one of those is a reliability KPI. The technician shortage is not adjacent to the reliability program — it is inside it.
The technology roadmap assumes a workforce that is, right now, walking out the door faster than it is being replaced.
The harder part: the knowledge, not just the headcount
If this were purely a headcount problem, it would be difficult but solvable — hire more, train faster. The harder problem is that what is retiring is not just labour. It is undocumented judgment.
The 58-year-old who knows Line 3 is not carrying a skill you can post a job description for. They are carrying twenty years of pattern recognition — the specific sound the compressor makes before it fails, the workaround for the sensor that reads high in summer, the reason nobody touches valve 14 during a changeover. None of that is in the CMMS. Most of it has never been written down. When they leave, it leaves with them, and the replacement starts from zero on a machine that has been running for two decades.
This is why the shortage is not a problem you can solve at the moment it becomes urgent. By the time the retirement happens, the window to capture the knowledge has already closed.
See the difference capture makes
The same retirement wave produces two very different outcomes depending on one variable: whether the knowledge was captured before the people left. Toggle between them.
What forward-looking operations are doing about it
There is no single fix, and anyone selling one is selling something. But the operations treating this seriously — ahead of the retirement wave, not during it — tend to be working several levers at once.
| Knowledge capture before departure | Structured capture of the undocumented judgment — shadowing, recorded troubleshooting walkthroughs, updating the CMMS with the why and not just the what — starting years before a key technician retires, not in their final month. |
| Apprenticeship & the trades pipeline | Rebuilding the pathway a generation was steered away from. The apprenticeship completion rate in Canada sits around 46 per cent — the pipeline leaks as much as it lacks intake. Both ends need work. |
| Making the work attractive to fill | The reqs that stay open longest are often the ones offering the least reason to take them. Shift structure, progression, and treating maintenance as a career rather than a cost centre all move the needle. |
| Technology as a multiplier, not a replacement | This is where predictive maintenance and AI genuinely help — not by removing technicians, but by making a smaller, less experienced team more effective. But it only works if there is a technician to act on the alert. Technology raises the ceiling; it does not fill the chair. |
| Cross-training for resilience | Single points of human failure are as real as single points of mechanical failure. The plants that weather a retirement well are the ones where more than one person understands each critical asset. |
The common thread: the shortage is a known, dated event, not a surprise. The retirement dates are on a spreadsheet in HR right now. The operations that do well are the ones treating those dates as scheduled maintenance on their own workforce — planned for, resourced, and started early.
How exposed is your team's knowledge?
Eight signals. Check what's true today — this reflects the resilience you already have in place, not a diagnosis.
What to watch next
The maintenance technician shortage is the rare crisis that arrives on a published schedule. We know roughly how many are leaving, and roughly when, because their retirement dates already exist.
That makes it, oddly, one of the more solvable problems in operations — not easy, but solvable — because it can be planned for. The plants that struggle will be the ones that treated it as a surprise. The plants that do well will be the ones that treated their own workforce the way they treat their critical equipment: knowing the failure is coming, and doing the unglamorous preventive work early enough to matter.
The sensor that predicts a bearing failure is impressive. But someone still has to change the bearing. The next decade of reliability depends less on the sophistication of the alert, and more on whether there is anyone left on the floor to answer it.
Planning knowledge capture ahead of your own retirement wave? Happy to compare notes.
Get in touch →Where these numbers come from
- Deloitte & The Manufacturing Institute, Manufacturers Need as Many as 3.8 Million New Employees by 2033
- Manufacturing Dive, Manufacturing could be short 1.9M workers if the talent gap isn't fixed
- JLL, Critical skilled-trades shortage threatens $1T in economic losses — workforce over-55 share
- RBC Economics, Powering Up: Preparing Canada's Skilled Trades — 700K retirements by 2028
- BuildForce Canada / regional trades outlook — Canadian manufacturing workforce age profile and apprenticeship completion rates
- Hero image: Pexels (free stock, commercial use)